X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/11c4ce5827158ef3f817f2536e220d95e0576e6e..fefd563815976d9f231bebc8182b44a7716de0c1:/youtube_dl/extractor/arte.py

diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py
index b061b95..e7a91a1 100644
--- a/youtube_dl/extractor/arte.py
+++ b/youtube_dl/extractor/arte.py
@@ -1,22 +1,31 @@
 import re
 import json
+import xml.etree.ElementTree
 
 from .common import InfoExtractor
 from ..utils import (
-    # This is used by the not implemented extractLiveStream method
-    compat_urllib_parse,
-
     ExtractorError,
     unified_strdate,
 )
 
 class ArteTvIE(InfoExtractor):
-    _VALID_URL = r'(?:http://)?www\.arte.tv/guide/(?:fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
+    """
+    There are two sources of video in arte.tv: videos.arte.tv and
+    www.arte.tv/guide, the extraction process is different for each one.
+    The videos expire in 7 days, so we can't add tests.
+    """
+    _EMISSION_URL = r'(?:http://)?www\.arte.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
+    _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?P<lang>fr|de)/.*-(?P<id>.*?).html'
     _LIVE_URL = r'index-[0-9]+\.html$'
 
     IE_NAME = u'arte.tv'
 
+    @classmethod
+    def suitable(cls, url):
+        return any(re.match(regex, url) for regex in (cls._EMISSION_URL, cls._VIDEOS_URL))
+
     # TODO implement Live Stream
+    # from ..utils import compat_urllib_parse
     # def extractLiveStream(self, url):
     #     video_lang = url.split('/')[-4]
     #     info = self.grep_webpage(
@@ -44,17 +53,27 @@ class ArteTvIE(InfoExtractor):
     #     video_url = u'%s/%s' % (info.get('url'), info.get('path'))
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        name = mobj.group('name')
-        # This is not a real id, it can be for example AJT for the news
-        # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
-        video_id = mobj.group('id')
+        mobj = re.match(self._EMISSION_URL, url)
+        if mobj is not None:
+            lang = mobj.group('lang')
+            # This is not a real id, it can be for example AJT for the news
+            # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
+            video_id = mobj.group('id')
+            return self._extract_emission(url, video_id, lang)
+
+        mobj = re.match(self._VIDEOS_URL, url)
+        if mobj is not None:
+            id = mobj.group('id')
+            lang = mobj.group('lang')
+            return self._extract_video(url, id, lang)
 
         if re.search(self._LIVE_URL, video_id) is not None:
             raise ExtractorError(u'Arte live streams are not yet supported, sorry')
             # self.extractLiveStream(url)
             # return
 
+    def _extract_emission(self, url, video_id, lang):
+        """Extract from www.arte.tv/guide"""
         webpage = self._download_webpage(url, video_id)
         json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
 
@@ -68,9 +87,20 @@ class ArteTvIE(InfoExtractor):
                      'description': player_info['VDE'],
                      'upload_date': unified_strdate(player_info['VDA'].split(' ')[0]),
                      'thumbnail': player_info['programImage'],
+                     'ext': 'flv',
                      }
 
         formats = player_info['VSR'].values()
+        def _match_lang(f):
+            # Return true if that format is in the language of the url
+            if lang == 'fr':
+                l = 'F'
+            elif lang == 'de':
+                l = 'A'
+            regexes = [r'VO?%s' % l, r'V%s-ST.' % l]
+            return any(re.match(r, f['versionCode']) for r in regexes)
+        # Some formats may not be in the same language as the url
+        formats = filter(_match_lang, formats)
         # We order the formats by quality
         formats = sorted(formats, key=lambda f: int(f['height']))
         # Pick the best quality
@@ -78,9 +108,38 @@ class ArteTvIE(InfoExtractor):
         if format_info['mediaType'] == u'rtmp':
             info_dict['url'] = format_info['streamer']
             info_dict['play_path'] = 'mp4:' + format_info['url']
-            info_dict['ext'] = 'mp4'
         else:
             info_dict['url'] = format_info['url']
-            info_dict['ext'] = 'mp4'
 
         return info_dict
+
+    def _extract_video(self, url, video_id, lang):
+        """Extract from videos.arte.tv"""
+        ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
+        ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
+        ref_xml = self._download_webpage(ref_xml_url, video_id, note=u'Downloading metadata')
+        ref_xml_doc = xml.etree.ElementTree.fromstring(ref_xml)
+        config_node = ref_xml_doc.find('.//video[@lang="%s"]' % lang)
+        config_xml_url = config_node.attrib['ref']
+        config_xml = self._download_webpage(config_xml_url, video_id, note=u'Downloading configuration')
+
+        video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
+        def _key(m):
+            quality = m.group('quality')
+            if quality == 'hd':
+                return 2
+            else:
+                return 1
+        # We pick the best quality
+        video_urls = sorted(video_urls, key=_key)
+        video_url = list(video_urls)[-1].group('url')
+        
+        title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
+        thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
+                                            config_xml, 'thumbnail')
+        return {'id': video_id,
+                'title': title,
+                'thumbnail': thumbnail,
+                'url': video_url,
+                'ext': 'flv',
+                }